3D convolutional neural networks for Alzheimer's disease diagnosis
Year: 2019
Type: dissertation
Abstract: Alzheimer's disease (AD) is an incurable neurodegenerative disease. Magnetic \nResonance Imaging (MRI) is used to assess the damage caused by the disease and to \ncapture complex changes on the brain. In this work, we study how 3D Convolutional \nNeural Networks exploit the 3D nature of the MRI volume and how they can predict AD \ndiagnosis. The neural network has been trained and evaluated in two different \nscenarios: a two way classification between AD and cognitive normal (CN), and a three \nway classification between CN, AD and an initial stage of the disease called Mild \nCognitive Impairment (MCI). Results classifying CN vs AD show a great performance \nusing two different image subsets and results for CN vs MCI vs AD show a nice \nperformance considering the lack of data and some data augmentation techniques that \naffect negatively to the model performance.
Author: Aitor Lucas Castellano
Language: English
Subfield: Psychiatry and Mental health
Field: Medicine
Domain: Health Sciences
Open Access status: green
OpenAlex ID: W2981168099